Sensors and Machine Learning Methods for Intent Detection in Lower-Limb Prosthetic Devices: A Review

Advances in lower-limb prosthetic (LLP) devices have greatly enhanced mobility for amputees, but achieving smooth, intuitive control remains a significant hurdle. A vital aspect of this challenge is the ability of prosthetic devices to detect user intentions, such as starting or stopping movement, changing locomotion modes, or adjusting speed. Reliable intent detection (ID) is crucial for improving adaptability, reducing the cognitive effort required from users, and maintaining safety, especially in changing environments. This systematic review, conducted in accordance with PRISMA guidelines, involved a comprehensive analysis of 60 full-text scientific articles. The review focuses on four critical areas: 1) how ID is defined across various studies; 2) the sensor technologies used and sensor placement; 3) the machine learning (ML) methods developed for ID; and 4) the metrics used to evaluate the performance of these methods. In summary, this review provides an in-depth look at the current developments in LLP ID, identifies critical gaps, and proposes areas for future research to enhance prosthesis functionality.